By Chisom Obiudo
Take the most recent strategic plan your board approved. It will be in the board pack. Now look for a second document: a written list of the assumptions on which the plan depends.
If those assumptions are absent from the strategic plan, risk analysis or minutes, the board will struggle to demonstrate that it tested them.
Discussion and testing are not the same. The difference becomes clear when an untested assumption fails.
Principle 4 of the King IV Code on Corporate Governance provides that the board should recognise that the organisation’s purpose, risks and opportunities, strategic plan, business model, performance and sustainable development are inseparable elements of value creation.
That is why a board cannot approve a strategic plan merely because its projections appear convincing.
It must test the assumptions, evidence, risks and warning signs on which the plan depends. Management may prepare the draft, but the board remains responsible for appraising, approving and monitoring it.
Why directors stay silent
The board’s duty to challenge the strategic plan is clear, yet individual directors may hesitate to do so.
Questioning a proposal backed by the chief executive may strain relationships, appear obstructive, or expose a director to being proved wrong in front of colleagues. The personal risk falls on one director, while the benefit belongs to the whole board.
Silence should therefore not be treated as agreement. Before the meeting, assign one director to present the strongest case against the strategic plan. This makes challenge an agreed board responsibility rather than a personal confrontation.
Three questions before the board votes
A structured challenge begins with three questions.
What is the strategic plan based on? Consider a proposed expansion into a neighbouring market. The plan may assume that the licence will be granted within the year, that a suitable distributor is available, that the exchange rate will remain stable, and that the existing management team can oversee operations in two countries.
Once written down, each assumption can be tested. If left unstated, a failed assumption may later become the reason the strategic plan fails.
What would have to be true for the strategic plan to succeed? Turn each assumption into a condition the board can observe and test. Record the supporting evidence, rate the board’s confidence as high, medium or low, and identify the earliest warning sign that the condition no longer holds. A board can monitor evidence and warning signs. It cannot monitor optimism.
What is the strongest argument against approving the strategic plan? Ask what a well-informed opponent would say about it. This encourages the board to consider weaknesses that management may not have anticipated, rather than limiting the discussion to objections that management has already prepared to answer.
Where AI helps and where it does not
AI can help the board prepare for the three questions by quickly identifying possible assumptions, risks and arguments against the strategic plan. It is not influenced by relationships with the chief executive or by the fear of appearing obstructive. This makes AI useful as a preparation tool, but not as an independent expert or decision-maker.
AI can only work with the information it receives. Unless relevant context is provided, it does not know the organisation’s market, financial position, regulatory obligations, risk appetite or history. It may generate objections that sound convincing but are irrelevant or incorrect. Treat its output as questions requiring evidence from management, not as proof that the strategic plan is sound or flawed. The judgement remains with the board.
Using AI to prepare the challenge
Run the exercise before the board pack is finalised so that management has time to gather the required evidence. The company secretary can coordinate the exercise with the executive responsible for the strategic plan and the team responsible for risk management. Use only an AI tool that the organisation has approved for this purpose.
Confidentiality comes first. If the tool is not approved for confidential information, do not upload or paste the strategic plan into it. Remove names, financial figures, counterparties, personal information, commercial terms and other non-public information. Describe the strategic plan only in general terms. Before using the tool, confirm how the AI vendor stores the information entered, who can access it, how long it is retained, and whether it is used to train the provider’s models.
The following prompt can then be adapted:
Act as a constructive critic assisting our board. Here is an anonymised description of the strategic plan: [describe the strategic plan]. Based only on the information provided, identify the assumptions on which the strategic plan depends, the evidence management should provide, the earliest warning signs, how the strategic plan could fail and the strongest argument against approval. Do not invent facts, figures, market information or sources. If important information is missing, state what additional information is required. End with no more than ten specific questions the board should answer before deciding.
